| """Synthetic text conversion-fidelity cases, not a quality benchmark.""" |
| CASES = [{'id': 'en_billing', |
| 'state': 'The customer was charged twice for a single order.', |
| 'question': {'type': 'choice', |
| 'instructions': 'Which team should handle this?', |
| 'criteria': {'billing': 'Payments and invoices', |
| 'shipping': 'Delivery and tracking', |
| 'account': 'Login and passwords'}}}, |
| {'id': 'en_receipt_yes', |
| 'state': 'The customer has a receipt.', |
| 'question': {'type': 'noul', 'instructions': 'Does the customer have a receipt?'}}, |
| {'id': 'en_receipt_no', |
| 'state': 'The customer lost the receipt and has no copy.', |
| 'question': {'type': 'noul', 'instructions': 'Does the customer have a receipt?'}}, |
| {'id': 'en_severity', |
| 'state': 'The service is unavailable for every user.', |
| 'question': {'type': 'score', |
| 'instructions': 'How severe is the outage?', |
| 'criteria': ['None', 'Minor', 'Major', 'Complete']}}, |
| {'id': 'es_account', |
| 'state': 'El usuario no puede iniciar sesión.', |
| 'question': {'type': 'choice', |
| 'instructions': '¿Qué equipo?', |
| 'criteria': {'cuenta': 'Acceso a cuenta', 'pago': 'Pagos', 'envío': 'Entregas'}}}, |
| {'id': 'zh_photo', |
| 'state': '用户说：请帮我拍一张照片。', |
| 'question': {'type': 'choice', |
| 'instructions': '应该选择哪项能力？', |
| 'criteria': {'摄影': '拍摄照片或视频', '写作': '撰写文章和代码', '闲聊': '一般问答'}}}, |
| {'id': 'zh_write', |
| 'state': '用户希望制作一份介绍太阳系的网页。', |
| 'question': {'type': 'choice', |
| 'instructions': '应该选择哪项能力？', |
| 'criteria': {'摄影': '拍摄照片或视频', '写作': '撰写文章和代码', '闲聊': '一般问答'}}}, |
| {'id': 'zh_complete', |
| 'state': '文件已经保存，应用返回成功，用户可以打开文件。', |
| 'question': {'type': 'noul', 'instructions': '任务是否已经完成？'}}, |
| {'id': 'zh_incomplete', |
| 'state': '已经开始保存，但还没有收到结果。', |
| 'question': {'type': 'noul', 'instructions': '任务是否已经完成？'}}, |
| {'id': 'zh_failure', |
| 'state': '机器人返回超时错误，动作未完成。', |
| 'question': {'type': 'choice', |
| 'instructions': '当前结果属于哪一种？', |
| 'criteria': {'成功': '已确认完成', '失败': '执行出现错误', '等待': '尚未返回结果'}}}, |
| {'id': 'zh_urgency', |
| 'state': '用户说随时都可以，今天不用处理。', |
| 'question': {'type': 'score', 'instructions': '紧急程度如何？', 'criteria': ['不紧急', '一般', '紧急', '必须立即处理']}}, |
| {'id': 'zh_ambiguous', |
| 'state': '用户说：帮我处理一下那个。没有其他上下文。', |
| 'question': {'type': 'choice', |
| 'instructions': '下一步应该是什么？', |
| 'criteria': {'询问': '获取缺少的信息', '拍照': '立即拍照', '写作': '立即写一篇文章'}}}, |
| {'id': 'zh_negation', |
| 'state': '不要拍照，我只是想知道摄像头是否可用。', |
| 'question': {'type': 'noul', 'instructions': '用户是否要求拍照？'}}, |
| {'id': 'zh_json', |
| 'state': {'工具': '保存文件', '状态': 'failed', '原因': '磁盘空间不足'}, |
| 'question': {'type': 'choice', |
| 'instructions': '如何分类？', |
| 'criteria': {'成功': None, '失败': None, '等待': None}}}, |
| {'id': 'en_ten_options', |
| 'state': 'A refund was promised yesterday.', |
| 'question': {'type': 'choice', |
| 'instructions': 'What topic?', |
| 'criteria': {'0': 'Billing', |
| '1': 'Delivery', |
| '2': 'Refund', |
| '3': 'Password', |
| '4': 'Support', |
| '5': 'Address', |
| '6': 'Fraud', |
| '7': 'Warranty', |
| '8': 'Subscription', |
| '9': 'Other'}}}, |
| {'id': 'one_option', |
| 'state': 'A single valid candidate.', |
| 'question': {'type': 'choice', |
| 'instructions': 'Choose the available option.', |
| 'criteria': {'only': 'The only candidate'}}}, |
| {'id': 'empty_state', |
| 'state': '', |
| 'question': {'type': 'noul', 'instructions': 'Is there enough information to confirm completion?'}}, |
| {'id': 'long_text', |
| 'state': 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund ', |
| 'question': {'type': 'choice', |
| 'instructions': 'What topic?', |
| 'criteria': {'refund': None, 'shipping': None}}}, |
| {'id': 'en_billing_reversed', |
| 'state': 'The customer was charged twice for a single order.', |
| 'question': {'type': 'choice', |
| 'instructions': 'Which team should handle this?', |
| 'criteria': {'account': 'Login and passwords', |
| 'shipping': 'Delivery and tracking', |
| 'billing': 'Payments and invoices'}}}, |
| {'id': 'es_account_reversed', |
| 'state': 'El usuario no puede iniciar sesión.', |
| 'question': {'type': 'choice', |
| 'instructions': '¿Qué equipo?', |
| 'criteria': {'envío': 'Entregas', 'pago': 'Pagos', 'cuenta': 'Acceso a cuenta'}}}, |
| {'id': 'zh_photo_reversed', |
| 'state': '用户说：请帮我拍一张照片。', |
| 'question': {'type': 'choice', |
| 'instructions': '应该选择哪项能力？', |
| 'criteria': {'闲聊': '一般问答', '写作': '撰写文章和代码', '摄影': '拍摄照片或视频'}}}, |
| {'id': 'zh_write_reversed', |
| 'state': '用户希望制作一份介绍太阳系的网页。', |
| 'question': {'type': 'choice', |
| 'instructions': '应该选择哪项能力？', |
| 'criteria': {'闲聊': '一般问答', '写作': '撰写文章和代码', '摄影': '拍摄照片或视频'}}}, |
| {'id': 'zh_failure_reversed', |
| 'state': '机器人返回超时错误，动作未完成。', |
| 'question': {'type': 'choice', |
| 'instructions': '当前结果属于哪一种？', |
| 'criteria': {'等待': '尚未返回结果', '失败': '执行出现错误', '成功': '已确认完成'}}}, |
| {'id': 'zh_ambiguous_reversed', |
| 'state': '用户说：帮我处理一下那个。没有其他上下文。', |
| 'question': {'type': 'choice', |
| 'instructions': '下一步应该是什么？', |
| 'criteria': {'写作': '立即写一篇文章', '拍照': '立即拍照', '询问': '获取缺少的信息'}}}, |
| {'id': 'zh_json_reversed', |
| 'state': {'工具': '保存文件', '状态': 'failed', '原因': '磁盘空间不足'}, |
| 'question': {'type': 'choice', |
| 'instructions': '如何分类？', |
| 'criteria': {'等待': None, '失败': None, '成功': None}}}, |
| {'id': 'en_ten_options_reversed', |
| 'state': 'A refund was promised yesterday.', |
| 'question': {'type': 'choice', |
| 'instructions': 'What topic?', |
| 'criteria': {'9': 'Other', |
| '8': 'Subscription', |
| '7': 'Warranty', |
| '6': 'Fraud', |
| '5': 'Address', |
| '4': 'Support', |
| '3': 'Password', |
| '2': 'Refund', |
| '1': 'Delivery', |
| '0': 'Billing'}}}, |
| {'id': 'long_text_reversed', |
| 'state': 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund refund ' |
| 'refund refund refund refund refund refund refund refund refund refund refund refund ', |
| 'question': {'type': 'choice', |
| 'instructions': 'What topic?', |
| 'criteria': {'shipping': None, 'refund': None}}}] |
| from duplex_cases import CASES as DUPLEX_CASES |
| CASES.extend(DUPLEX_CASES) |
|
|
